Skip to content

Baprojet

Baprojet yenye ilichaguliwa

Ba système ya recherche, contribution open source na ba outil ya elimu mu domaine tofauti ya AI ya sasa.

  • STEM education platform

    DarAkili

    A multilingual STEM education platform where African professionals share free mini-courses in the languages their communities use, connecting expertise across the diaspora and the continent.

    STEM Education · African Languages · Community · Platform

    DémoYouTube

  • NAACL 2025 · Long Papers

    IrokoBench

    A human-translated benchmark for evaluating natural-language inference, mathematical reasoning, and knowledge-based question answering across 17 African languages.

    LLM Evaluation · African Languages · Reasoning · Multilingual NLP

    Article

  • ACL 2025 · Long Papers

    INJONGO

    A culturally grounded, open benchmark for intent detection and slot filling, created with native-speaker utterances across 16 African languages.

    Intent Detection · Slot Filling · African Languages · LLM Evaluation

    Article

  • Mozilla Open Source Support Award

    Masakhane Web Platform

    A Mozilla-funded web platform that makes community-built translation models for African languages available through an accessible interface.

    NLP · Machine Translation · African Languages · Web

    Démo

  • EMNLP Findings 2020 · Community research

    Masakhane Initiative

    A participatory research community where African researchers collaborate on datasets, models, and methods for African-language technology. Masakhane means "We Build Together" in isiZulu.

    NLP · African Languages · Community · Research

    GitHubDémo

  • ICADL 2021 Best Paper Award

    ORKG Leaderboards

    Engineered the Leaderboards feature for the Open Research Knowledge Graph (ORKG) to automatically track and visualize state-of-the-art progress across scientific publications using Knowledge Graphs.

    Knowledge Graphs · NLP · Scholarly IE · Python

    GitHubArticle

  • NeurIPS 2021 Black in AI Workshop

    LiSTra Speech Translation

    The first English-to-Lingala automatic speech translation dataset and baseline. Released a full pipeline for constructing similar datasets in other low-resource languages.

    ASR · Speech Translation · Lingala · Low-Resource NLP

    GitHubArticle